The Canadian Internet Use Survey (CIUS) measures household access to the Internet and individual online behaviours including electronic commerce.
Data release – October 28, 2013 (Individual Internet use and E-commerce); November 26, 2013 (Household component).
The Canadian Internet Use Survey (CIUS) was redesigned in 2010 to better measure the type and speed of household Internet connections. It is a hybrid survey that measures both household Internet access and the individual online behaviours of a selected household member. It replaces the previous CIUS, a biennial survey conducted in 2005, 2007 and 2009. As the new survey has two distinct components - household and individual - with revised and streamlined questions, it is not appropriate to directly compare results from the two surveys in most cases.
The CIUS measures the availability, type and speed of home Internet connections. It then selects a member from the household to measure the extent and scope of online behaviour including the location, frequency and intensity of use, specific uses including the purchase of products and services (electronic commerce), and other related issues such as online security practices. This content is supplemented by individual and household characteristics (e.g., age, income, family type) and some geographic detail (e.g. province and Census Metropolitan Area).
The CIUS results are used by the federal government to build and evaluate policies and programs related to uptake and barriers to use, high speed access and electronic commerce. International agencies such as the Organization for Economic Cooperation and Development (OECD) use the results for benchmarking and comparison studies. The CIUS data support research initiatives and micro data are made available to universities under the Data Liberation Initiative. Estimates are used by the private sector for market research and for consultation on Internet-related regulatory issues. Finally, CIUS findings are widely quoted in the media, given the interest in the Internet and its users.
The target population is residents of Canada 16 years of age or older excluding: Residents of the Yukon, Northwest Territories and Nunavut, Inmates of Institutions, Persons living on Indian Reserves, and Full time members of the Canadian Forces.
This is a sample survey with a cross-sectional design.
The final sample sizes were 30,817 for the Household component and 22,615 for the Individual component.
The CIUS was administered to a sub-sample of the individuals already selected for the Labour Force Survey (LFS), record number 3701. The LFS uses a stratified, multi-stage cluster design employing probability sampling at all stages to select a representative sample of households (see Target Population). Every month, one-sixth of the LFS sample is replaced by a new "birth" panel of dwellings.
The 2012 CIUS started with the households in the non-birth LFS panels, 3 in October and 3 in November. For 2 panels each month, the Household component was administered to all eligible households with the computer application randomly selecting one member 16 years of age or older, for the Individual component. To oversample, just the Household component was administered to rural households in the 3rd LFS panel each month.
Data collection for this reference period: 2012-10-14 – 2012-11-26
Responding to this survey is voluntary.
Data are collected directly from survey respondents.
The LFS interview was completed by a responsible member of the household who generally provides the LFS responses for all members. Following the LFS interview, and subject to operational constraints, the interviewer then requested that this same member answer the household component questions of the CIUS. Following the CIUS Household component portion, the interviewer will then ask to speak to the randomly selected member for the CIUS Individual component.
Depending on availability, some of the CIUS Individual component interviews were attempted immediately following the Household component. Since the CIUS Individual component is non-proxy however, only the household member randomly selected can answer the questions. Interviewers routinely ask for and record the best time to call back in order to complete the interview and the automated call scheduler manages follow-up calls in order to try different times of day throughout the collection period.
Interviews were conducted from Statistics Canada's regional offices using a Computer Assisted Telephone Interviewing (CATI) application. To reduce collection costs, no Computer Assisted Personal Interviewing (CAPI) was conducted and this was treated by adjusting the weights using reported LFS characteristics (see Estimation).
The questionnaire incorporates several features to ensure high quality. For example, multiple edits in the CATI application assess reported data with unusual values for incomes and on-line expenditures. For most of these failures, the interviewer has the ability to override the edit failure if it cannot be resolved. Other edits check for logical inconsistencies and prompt the interviewer to correct responses with the help of the respondent. The interviewer has the ability to enter a response of Don't Know or Refused if the respondent does not answer the question.
Once the data are collected, an extensive series of processing steps is undertaken including a top-down flow edit to correct questionnaire paths mistakenly followed. The editing and imputation processing identify inconsistent or missing items, and correct errors.
Suspiciously large reported values for income and the number or value of Internet orders were identified as "outliers" and treated by replacing the suspicious values by ones from respondents with similar characteristics (see Imputation).
Imputation is the process that supplies valid values for those responses that have been identified as either invalid or missing on the data file. The new values are supplied in such a way as to preserve the underlying structure of the data and to ensure that the resulting records will pass all required edits. Imputation was done using a nearest-neighbour method which searches for "donor" records from individuals with complete and consistent values. The recipient records are imputed by a donor chosen from a group of records with similar demographic characteristics.
CIUS imputation was limited to item non-response for household income and for the number and value of online orders. In the case of income, LFS data on the usual weekly earnings for a recipient were used to select the donor with the most similar values. In total, about 20,000 respondents (66%) reported their household income. Respondents who did not provide a dollar estimate of their income were asked for an income range. About 4,000 respondents (13%) did not provide any information on their income. The reported income ranges and the missing income information were imputed by donor values in a series of steps, depending on the available information.
For the number of online orders and their value, donor imputation was also used. Again, the donor records were chosen from a group of records with similar demographic characteristics, as well as, similar Internet shopping behaviour (e.g. types of goods and services). The relative imputation rate serves as a data quality indicator. The rates for 2012, based on value-weighted estimates are the following: for every 100 online orders estimated from the survey, about 12 orders were imputed. Similarly with the value of orders, for every $100 in online orders estimated from the survey, $10.80 was imputed.
Estimates are produced using weights attached to each sampled unit (i.e. household or individual). The weight of a sampled unit indicates the number of units in the population that the unit represents. The initial weight was provided by the LFS and incorporated the probability of selecting the unit in the sample, as well as other adjustments such as the treatment of non-response to the LFS. For CIUS, two sets of weights were produced: household weights for household-level estimates, and person weights for individual-level estimates.
The household weights for household-level estimates were produced in several steps.
A first adjustment was made to the initial weights to reflect the number of LFS panels used for CIUS. A second adjustment took into account the oversampled rural households. A third adjustment inflated the weights to represent the LFS CAPI respondents not interviewed for the CIUS. A fourth adjustment inflated the weights to represent the household-level non-respondents in the CIUS. The fifth adjustment involved calibrating the weights to demographic counts in two stages: the weights were first calibrated to provincial age group counts and some CMA counts, and then calibrated to provincial household size counts. The final CIUS household weight is the outcome of these five adjustments to the initial LFS sub-weight.
The person weights for individual-level estimates were produced in several steps as well. A first adjustment was made to the initial weights to reflect the number of LFS panels used for CIUS. A second adjustment inflated the weights to represent the LFS CAPI respondents not interviewed for the CIUS. A third adjustment inflated the weights to represent the household-level non-respondents in the CIUS. A fourth adjustment was done to account for the selection of a single household member, 16 years of age or older. A fifth adjustment inflated the weights to represent the person-level non-respondents. The sixth adjustment involved calibrating the weights to provincial age-sex demographic counts and some CMA counts. The final CIUS person weight is the outcome of these six adjustments to the initial LFS sub-weight.
The quality of the estimates is assessed using estimates of their coefficient of variation (CV). Given the complexity of the CIUS design, CVs cannot be calculated using a simple formula. Bootstrap replicate weights were used to establish the CVs of the estimates.
A comparison of social and demographic domains from CIUS was made with the previous survey to ensure consistency. Subject matter experts made selective data confrontations with data from other sources (e.g. the Survey of Household Spending, record number 3508), regulatory agencies (e.g. the Canadian Radio-television and Telecommunications Commission) and international organizations (e.g. the OECD).
Statistics Canada is prohibited by law from releasing any information it collects which could identify any person, business, or organization, unless consent has been given by the respondent or as permitted by the Statistics Act. Various confidentiality rules are applied to all data that are released or published to prevent the publication or disclosure of any information deemed confidential. If necessary, data are suppressed to prevent direct or residual disclosure of identifiable data.
Users requiring access to information may purchase custom tabulations. Estimates will be released subject to meeting the guidelines for analysis and maintenance of confidentiality.
The CIUS collection response rate was 81% for the Household component and 69% for the Individual component, with respect to the LFS households eligible to participate in the CIUS. The CIUS estimation response rate for the Household component and for the Individual component was 71% and 60% respectively, taking into account the CIUS non-response, the LFS non-response and the LFS CAPI cases excluded from CIUS (see Data sources).
The results estimated from CIUS are based on a sample of Canadians. The results obtained from asking the same questions of all Canadians would differ to some known extent. The extent of this sampling error is quantified by the coefficient of variation (CV) with the following guidelines:
- 16.5% and below Acceptable;
- 16.6% to 33.3% Marginal, with cautionary note; and
- Above 33.3% Unacceptable estimate.
Estimates that do not meet an acceptable level of quality are either flagged for caution or suppressed.
CVs are provided for the key survey variable: CU_Q01 Did you use the Internet during the past 12 months for personal use?